SENSING FOR MOBILE OBJECTS - Managing and Mining Sensor Data - page 300

In this chapter, we introduce the types of problems that arise from
managing spatiotemporal data and survey the recent research that ad-
dresses these issues. In our review, we attempt to cover work on data
management, object tracking (processing updates of moving objects),
and mining spatiotemporal data. Our aim in this chapter is twofold: (i)
to provide a review of the recent developments in each of the different ar-
eas of study relating to spatiotemporal data and (ii) to introduce work
on tracking and show how it relates to the database-centric research
(e.g. querying, indexing, and mining). While we attempt to provide
a broad overview of recent work in all of the mentioned areas, we pay
special attention to work which explicitly manages uncertainty in the
spatiotemporal data.
The rest of the chapter is organized as follows: first we will review
work on data management for spatiotemporal data, including indexing
and querying, in section 2. Next, we introduce the problem of tracking
and review some core and recent developments in that area in section 3.
In section 4, we review some recent work on mining spatiotemporal data
in three broad categories: (i) clustering, (ii) popular route discovery, and
(iii) identifying mobility patterns. We conclude by discussing directions
for future research.
2. Data Management for Mobile Objects
Database management systems for spatiotemporal data can be char-
acterized as one of two types: spatiotemporal database systems (STDB)
and moving object database systems (MOD). Both are used to manage
data collected from mobile objects, however, the specific problems each
solves is quite different. STDBs store the complete historic trajectory of
each object, and thus allows users to answer complex queries about user
movement over time . For example, “find all users that passed through
region A between 8 - 10am, through region B between 1 - 2pm and re-
gion C between 4 - 7pm”. In contrast, MODs only maintain the current
position of each object along with each object's velocity or heading infor-
mation if it is available. Therefore, while STDBs contain more complete
information, MODs provide access to location data that is consistently
current (see figure 2 ) . MODs are well suited to answer queries about the
current (and near-future) configuration of mobile objects. Both types of
database systems present a unique set of challenges due to the data they
manage. In this section, we introduce both systems, as well as some re-
cent work which addresses the challenges in ecient indexing and query
processing. In addition to the general issues with managing spatiotem-